CS 425 / ECE 428 Distributed Systems

Fall 2020Indranil Gupta (Indy)

August 25 – December 8, 2020Lecture 1-29

Web: courses.engr.illinois.edu/cs425/ All slides © IG

We’ve Made it very far!

Congratulations to everyone who’s made it so far in the course!

It has been a challenging year (to say the least) for everyone.

Our goal for today: see how far we’ve learnt on the topic of Distributed Systems.

2

Our First Goal in this Course was…

To Define the Term Distributed System(First lecture slide)

3

• Client-Server (NFS)• The Web• The Internet• A wireless network• DNS• Gnutella or BitTorrent (peer to peer overlays)• A “cloud”, e.g., Amazon EC2/S3, Microsoft Azure• A datacenter, e.g., NCSA, a Google datacenter, The Planet

Can you name some examples of Distributed Systems?

(First lecture slide)

4

What is a Distributed System?

(First lecture slide)

5

FOLDOC definition

A collection of (probably heterogeneous) automata whose distribution is transparent to the user so that the system appears as one local machine. This is in contrast to a network, where the user is aware that there are several machines, and their location, storage replication, load balancing and functionality is not transparent. Distributed systems usually use some kind of client-server organization.

(First lecture slide)

6

Textbook definitions

• A distributed system is a collection of independent computers that appear to the users of the system as a single computer. [Andrew Tanenbaum]

• A distributed system is several computers doing something together. Thus, a distributed system has three primary characteristics: multiple computers, interconnections, and shared state.[Michael Schroeder]

(First lecture slide)

7

A working definition for us

A distributed system is a collection of entities, each of which is autonomous, programmable, asynchronous and failure-prone, and which communicate through an unreliablecommunication medium.

• Entity=a process on a device (PC, PDA)• Communication Medium=Wired or wireless network• Our interest in distributed systems involves

– design and implementation, maintenance, algorithmics

• What Evidence/Examples have we seen?

(First lecture slide)

8

Problems we have seen since then• Time and Synchronization• Global States and Snapshots• Failure Detectors • Multicast• Mutual Exclusion • Leader Election • Consensus and Paxos• Gossiping• Peer to peer systems – Napster, Gnutella

Chord, BitTorrent• Cloud Computing and Hadoop• Sensor Networks• Structure of Networks• Datacenter Disaster Case Studies

Basic Theoretical Concepts

Cloud Computing

What Lies Beneath

9

• RPCs & Distributed Objects • Concurrency Control• 2PC and Paxos• Replication Control• Key-value and NoSQL stores• Stream Processing• Graph processing• Scheduling• Distributed File Systems • Distributed Shared Memory • Security

Problems we have seen since then (2)

Basic Building BlocksDistributed Services

(e.g., storage)

Cloud Computing

Old but Important (Re-emerging) 10

What This Course is About

• US Elections• Movies• Travel to Mars• Job Interviews• (Not Kidding)

11

What This Course is About

• US Elections: HW1• Movies: HW2• Travel to Mars: HW3• Job Interviews: HW4• (Not Kidding)

12

• Midterm• HW’s and MP’s

– (4 cr and Coursera) You’ve built a new cloud computing system from scratch!

– And beaten a state of the art system!

How to get good grades (and regrades, and jobs in some cases)

(& that standard devs are important!)

How far is your design from a full-fledged system?

Can you convince developers to use yourMapleJuice instead of Hadoop?

13

What This Course is About (2)

Rejoinder: Typical Distributed Systems Design Goals

• Common Goals:– Heterogeneity– Robustness– Availability– Transparency– Concurrency– Efficiency– Scalability– Security– Openness

(First lecture slide)

Do they make sense now?

14

Rejoinder: Typical Distributed Systems Design Goals

• Common Goals:– Heterogeneity – can the system handle a large variety of types of PCs and devices?– Robustness – is the system resilient to host crashes and failures, and to the

network dropping messages? – Availability – are data+services always there for clients?– Transparency – can the system hide its internal workings from the users?– Concurrency – can the server handle multiple clients simultaneously?– Efficiency – is the service fast enough? Does it utilize 100% of all resources?– Scalability – can it handle 100 million nodes without degrading service?

(nodes=clients and/or servers) How about 6 B? More?– Security – can the system withstand hacker attacks?– Openness – is the system extensible?– (Also: consistency, CAP, partition-tolerance, ACID, BASE, and others … )

(First lecture slide)

15

Problems we have seen in Class (and their relation to other courses)

• Time and Synchronization• Global States and Snapshots• Failure Detectors• Multicast Communications • Mutual Exclusion • Leader Election • Consensus and Paxos• Gossiping• Peer to peer systems – Napster, Gnutella

Chord• Cloud Computing • Sensor Networks• Structure of Networks• Datacenter Disaster Case Studies

Core Material of this course

Related to other graduate classes in

department (e.g., CS523, CS525,CS 498ISE, CS598WSI)

16

• RPCs & Distributed Objects • Concurrency Control• 2PC and Paxos• Replication Control• Key-value and NoSQL stores• Stream Processing• Graph processing• Scheduling• Distributed File Systems • Distributed Shared Memory • Security

Problems we have seen in Class (and their relation to other courses)

Core Material of this course

Related to CS 411/CS 511

Related to CS 523/561

Related to CS 421/CS 433

Related to CS 525

17

CS525: Advanced Distributed Systems(taught by Indy)

CS 525, Offered Spring 2021!– Looks at hot topics of research in distributed systems: cutting-edge papers on clouds+datacenters, p2p,

distributed machine learning, sensor/IoT networks, distributed algorithms, and other distributed systems– We will read many papers (and webpages) for cutting-edge systems (research and production)– If you liked CS425’s material, it’s likely you’ll enjoy CS525!– Project: Choose between Research project or Entrepreneurial project

• Research Project: Your project will build a cutting-edge research distributed system, and write and publish a paper on it

• Entrepreneurial Project: Your project will build a distributed system for a new startup company idea (your own!) and perform associated research with it

• Projects are in groups of your choosing (2-3).– Both graduates and undergraduates welcome! (UG fill this out for consent:

https://my.cs.illinois.edu/ugradrecs/petitions/ ).– Class size is around 70-100– Previous research projects published in journals and conferences, some great startup ideas too! 18

https://my.cs.illinois.edu/ugradrecs/petitions/

Other Related Grad Courses

• CS525 – Indy (next offered SP 2021)• CS598 CAL – Consensus, Blockchain (Ling Ren)• CS523 – Tianyin Xu• IoT classes: CS 598 WSI (Deepak Vasisht), CS 598 ISE (Matt Caesar)

• See also courses by Radhika Mittal (ECE, distributed storage), Andrew Miller (ECE, blockchain)

Questions?

A working definition for us

A distributed system is a collection of entities, each of which is autonomous, programmable, asynchronous and failure-prone, and which communicate through an unreliablecommunication medium.

[Is this definition still ok, or would you want to change it?]Think about it!

(First lecture slide)

21

Final Exam• Office Hours: Regular [All TAs and Indy] until final exam window starts

(usual schedule).– Exceptions posted on Piazza (check before heading out to an OH)

• Final Exam Window: See website– Syllabus: Includes all material since the start of the course. There may be more

emphasis on material since midterm.• Please check Piazza before (and during) finals: updates will be posted there

22

Course Evaluations• Main purpose: to give us feedback on how useful this course was to you

(and to improve future versions of the course)• I won’t see these evaluations until after you see your grades• Answer all questions• Please write your detailed feedback – this is valuable for future versions

of the course!

23